GYB Crash during refresh - ConnectionAbortedError: [WinError 10053]

[KeithR]

I am backing up a personal gmail.com account to the local drive.

After running perfectly since installation on 28 April, GYB 1.50 now quits during the refresh process with the trace copied below.  The sqlite file was not touched during the refresh cycle since it has the same time stamp as at did when "GYB needs to refresh..." was typed.  The crash happens at about the time it reaches 19,000 messages, perhaps the first message that needs a change?

The time to refresh 100 messages was increasing exponentially.  Initially it was a second or two.  By 18000 messages, it was out to 4 minutes.

What should I look at/for?

I ran a --fast-incremental backup to skip the refresh step which successfully backed up another 10 messages, then tried the full backup again.  

Trace:

GYB needs to refresh 64952 messages

Traceback (most recent call last):

  File "gyb.py", line 2429, in <module>

  File "gyb.py", line 1849, in main

  File "gyb.py", line 712, in callGAPI

  File "googleapiclient\_helpers.py", line 134, in positional_wrapper

  File "googleapiclient\http.py", line 1598, in execute

  File "gyb.py", line 1524, in refresh_message

  File "gyb.py", line 1449, in labelIdsToLabels

  File "gyb.py", line 712, in callGAPI

  File "googleapiclient\_helpers.py", line 134, in positional_wrapper

  File "googleapiclient\http.py", line 920, in execute

  File "googleapiclient\http.py", line 222, in _retry_request

  File "googleapiclient\http.py", line 191, in _retry_request

  File "gyb.py", line 107, in wrapped_request_method

  File "google_auth_httplib2.py", line 218, in request

  File "httplib2\__init__.py", line 1708, in request

  File "httplib2\__init__.py", line 1424, in _request

  File "httplib2\__init__.py", line 1376, in _conn_request

  File "http\client.py", line 1345, in getresponse

  File "http\client.py", line 307, in begin

  File "http\client.py", line 268, in _read_status

  File "socket.py", line 704, in readinto

  File "ssl.py", line 1241, in recv_into

  File "ssl.py", line 1099, in read

ConnectionAbortedError: [WinError 10053] An established connection was aborted by the software in your host machine

[16484] Failed to execute script gyb

[KeithR]

After running the --fast-incremental run, the full backup has got to Refreshed 31900 without crashing (after 7.5 hours).  Since it's approaching bedtime here, I'll update tomorrow morning.

If this fixed it, I'd like to know why!

[KeithR]

Not fixed.  Morning showed exactly the same trace as above.  I can't tell how far it got since the trace overwrites the "Refreshed xxx messages" line.

Help!

[Jay Lee]

The error indicates GYB's network connection was interrupted by something else on your machine. In my experience this is most often the Antivirus / security software you also have installed. Try disabling said software or exempting GYB.

Jay

On Tue, May 25, 2021, 6:48 PM KeithR <ke4...@gmail.com> wrote:

Not fixed.  Morning showed exactly the same trace as above.  I can't tell how far it got since the trace overwrites the "Refreshed xxx messages" line.

Help!

--
--
You received this message because you are subscribed to the Google
Groups "Got Your Back: Gmail Backup" group.
To post to this group, send email to got-yo...@googlegroups.com
To unsubscribe from this group, send email to
got-your-bac...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/got-your-back?hl=en?hl=en

---
You received this message because you are subscribed to the Google Groups "Got Your Back: Gmail Backup" group.
To unsubscribe from this group and stop receiving emails from it, send an email to got-your-bac...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/got-your-back/1209b7c2-094f-4bd8-99cf-699dc6637370n%40googlegroups.com.

[Keith Robards]

I disabled my Antivirus real time protection component and Windows Defender Firewall, but no change.  I've now added gyb.exe to the list of apps allowed through the Firewall and re-enabled both and running it again.

Is it normal that the refresh should slow down to such a low rate?  It used to run all the way through at the initial speed of a second or two for every 100 refreshes.  I experimented with a sizing script that added a label to every message to record progress and ever since then the refresh has taken many hours even though I've removed all those labels.  It now takes around 13 hours to refresh all 65,000 messages although the first 15,000 are done in 15 minutes.

[KeithR]

Same result.  Crashed after about 11 hours and refreshing over 50,000.  It doesn't seem reasonable that any protection software would allow this to run for 11 hours then kick it off.  Any other ideas?

[Jay Lee]

I'd suggest starting with a fresh backup and seeing where that gets you.

Jay

--

--
You received this message because you are subscribed to the Google
Groups "Got Your Back: Gmail Backup" group.
To post to this group, send email to got-yo...@googlegroups.com
To unsubscribe from this group, send email to
got-your-bac...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/got-your-back?hl=en?hl=en

---
You received this message because you are subscribed to the Google Groups "Got Your Back: Gmail Backup" group.
To unsubscribe from this group and stop receiving emails from it, send an email to got-your-bac...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/got-your-back/7ffddf2f-17b7-4137-9bb8-3178536c11a0n%40googlegroups.com.

[KeithR]

New backup has now been running for 11 hours and reached 49,000 of 65,059 messages.  It has shown this error three times:

<HttpError 500 when requesting https://gmail.googleapis.com/gmail/v1/users/me/messages/13920d6ee27d77f0?format=raw&fields=id%2ClabelIds%2CinternalDate%2Craw&alt=json returned "Internal error encountered.". Details: "[{'message': 'Internal error encountered.', 'domain': 'global', 'reason': 'backendError'}]">

Is that an error that might interfere with the quality of the backup?

[KeithR]

Minutes after I posted that, I got the same old crash.  I think my installation is kaput.

Next steps:

1. Running --fast-incremental to [hopefully] ensure the new backup gets completed
2. Will run --search after:2020-12-31 to get this year's messages refreshed
3. Will run --search before:2021-01-01 to get older messages refreshed (I will probably have to experiment with the date to get a refresh to run in the time before it crashes)

Since I don't normally relabel older messages, I can probably get away with using #2 as my regular backup.